Nobel Economist Acemoglu: AI Will Add Just 0.55% Productivity

The economist who won the 2024 Nobel Prize for his work on institutions and prosperity has a stark message for AI optimists: the technology will deliver a fraction of the economic transformation Wall Street expects, while accelerating dangerous concentrations of corporate power.

Daron Acemoglu of MIT estimates AI will generate roughly 0.55% in total factor productivity gains over the next decade. He projects only about 5% of tasks will be profitably automated in the near term, translating to perhaps a 1% to 1.5% increase in GDP. These figures stand in sharp contrast to the euphoric forecasts driving tech valuations.

When asked how much of the current AI discourse he finds intellectually serious, Acemoglu didn't hesitate: about 20%. The remaining 80%, he told Fortune in a recent interview, ranges from speculative to nearly fictional.

Why it matters

Acemoglu's framework treats AI not as an isolated technology question but as a stress test for democratic institutions and economic inclusion. His concern centers on whether AI systems broaden participation and reward innovation, or whether they concentrate power and extract value at scale. The answer shapes not just productivity statistics but the stability of liberal democracies facing a generation of workers who may find their credentials suddenly devalued.

The extractive model

Acemoglu rejects the term "capitalism" as analytically useless, arguing it lumps together Sweden, Egypt, Argentina, and the United States under one label despite radically different institutional arrangements. His preferred distinction, developed across books including Why Nations Fail, separates inclusive institutions from extractive ones.

Today's AI hyperscalers, he argues, fit the extractive pattern: concentrated ownership, regulatory capture, and business models that extract data and attention at industrial scale. "We should be talking about the enormous increase in corporate power and monopoly," he said. "What we should be talking about is the displacement and unequalizing roles of AI."

The productivity gap

Acemoglu's skepticism about AI's economic upside rests on decades of studying automation waves. Productivity gains only materialize when machines perform tasks significantly cheaper or better than humans. Marginal improvements or high integration costs erase the advantage, even when automation spreads widely.

Most research on AI productivity, he notes, focuses on easy, well-defined tasks where context is clear. These aren't representative of the broader economy, and current AI systems struggle with complex, high-stakes professional work that requires reading a room or connecting non-obvious dots across domains.

For the massive productivity gains that AI bulls project, "we really, really need something close to AGI," Acemoglu said, referring to artificial general intelligence. He's skeptical we're close.

The revolution risk

Acemoglu offers a paradoxical warning to business leaders: they should hope he's right that AI won't be transformative. If 30% to 40% of new university graduates can't find jobs, he asks, what happens to democracy and social peace? Historically, that scenario produces revolutions.

A generation of workers who trained for an economy that AI has since restructured represents a constituency that has never stayed quiet. Social media adds an unpredictable variable that history offers no guide to navigating.

What would fix it

The United States needs a genuine conversation about what is socially desirable from AI, not just what is technically possible or financially profitable for a handful of companies. That conversation must center on wages, jobs, shared prosperity, and meaningful work.

It also requires serious global governance, including cooperation with China, which Acemoglu says is ahead of the U.S. in integrating AI into manufacturing and commerce. The current geopolitical climate makes that nearly impossible. "The only bipartisan issue in the United States right now is China bashing," he noted.

The deeper failure, in his view, is one of imagination: an inability to articulate what a genuinely human-centered AI future would look like and the political will to demand it. "We're all so blindly taken in by what OpenAI, Anthropic, and a few other hyperscalers are offering," he said, "because we haven't articulated a reasonable alternative."

These details were first reported by Fortune.